The purpose of this fork is to add some pronunciation evaluation related feature extraction in pocketsphinx.js. Look under src/
folder to check:
featex.h
,featex.cpp
,psRecognizer.h
andpsRecognizer.cpp
for relevant modifications. This work is done under GSoC 2017: Pronunciation Evaluation project and mentored by James Salsman. The original source code for feature extraction written as a standalone C binary can be found at https://github.com/jsalsman/featex.
For details check out posts at: https://pronunce.blogspot.in/
PocketSphinx.js is a speech recognizer that runs entirely in the web browser. It is built on:
- a speech recognizer written in C (PocketSphinx) converted into JavaScript using Emscripten,
- an audio recorder using the web audio API. The audio recorder can be used independently to build other kinds of audio-related web applications. There is a more detailed documentation in
doc/AudioRecorder/README.md
.
You can try it on the project page: https://syl22-00.github.io/pocketsphinx.js and have a look at our FAQ.
Table of contents:
- Overview
- Compilation of
pocketsphinx.js
- API of
pocketsphinx.js
- Using
pocketsphinx.js
inside a Web Worker withrecognizer.js
- Wiring
recognizer.js
to the audio recorder - Live demo
- Test suite
- Notes about speech recognition and performance
- License
This project includes several components that can be used independently:
pocketsphinx.js
, a JavaScript library generated by emscripten which is basically PocketSphinx wrapped to provide a simpler API, and compiled into JavaScript.recognizer.js
, a wrapper aroundpocketsphinx.js
inside a Web Worker to unload the UI thread from downloading and running the large JavaScript file and running the costly speech recognition process.audioRecorder.js
, an audio recording library, based on Recorderjs. It converts the recorded samples to the proper sample rate and passes them to the recognizer. There is a more detailed documentation indoc/AudioRecorder/README.md
.callbackManager.js
, a small utility to interact with Web Workers with calls and callbacks rather than message passing.
The file webapp/live.html
illustrates how these work together in a real application, that is a good starting point. Make sure you load it through a web server or start Chrome with --disable-web-security
. For instance, you can start a small web server with python -m SimpleHTTPServer
in the base directory and open https://localhost:8000/webapp/live.html
in your browser.
Also Note that on very recent versions of Chrome, the app must be served through https (or localhost, or with --disable-web-security
) to allow audio recording.
There is also a live demo for Chinese. To try it, open https://localhost:8000/webapp/live_zh.html
in your browser.
In addition to speech recognition, there is also a keyword spotting functionality that detects a specific word or phrase in the audio input. There are live demos in webapp/live.html
(via a key phrases file) and webapp/live_kws.html
(via keyword spotting API).
A prebuilt version of pocketsphinx.js
is available in webapp/js
, or you can build it yourself. Below is the procedure on Linux (and Mac OS X). On Windows, refer to the emscripten manual.
The code of upstream Pocketsphinx
is referenced as git submodules, so before you compile, you must make sure you have retrieved the content of the submodules. You can either clone pocketsphinx.js
with the --recursive
option, or run:
$ git submodule init
$ git submodule update
You will need:
- emscripten (which implies also node.js and LLVM-fastcomp compiler, see emscripten docs for instructions on how to get it),
- CMake.
The build is a classic CMake cross-compilation, using the toolchain provided by emscripten:
$ cd .../pocketsphinx.js # This folder
$ mkdir build
$ cd build
$ cmake -DEMSCRIPTEN=1 -DCMAKE_TOOLCHAIN_FILE=../../emsdk-portable/emscripten/1.37.16/cmake/Modules/Platform/Emscripten.cmake ..
$ make
This generates pocketsphinx.js
. At this point, optimization level and other compilation parameters are hard-coded, so modify CMakeLists.txt
directly if you would like to change them.
The compilation process can package the acoustic models inside the resulting JavaScript file and also, possibly, language models and dictionary files. It can also produce a JavaScript file that does not include these files, and load their content using separate files. The later being necessary if the files are large. If you would like to package your own models, you should specify where they are when running cmake
. For that, place all models you want to package inside a base folder and specify the files or sub-folders you want to include.
For instance, to package acoustic models, place them inside a HMM_BASE
folder. Each model being in its own folder inside HMM_BASE
:
$ cmake -DEMSCRIPTEN=1 -DCMAKE_TOOLCHAIN_FILE=path_to_emscripten/cmake/Modules/Platform/Emscripten.cmake -DHMM_BASE=/path/to/models -DHMM_FOLDERS="model1;model2;..." ..
If you only need to package one model, you can also do:
$ cmake -DEMSCRIPTEN=1 -DCMAKE_TOOLCHAIN_FILE=path_to_emscripten/cmake/Modules/Platform/Emscripten.cmake -DHMM_BASE=/path/to/models -DHMM_FOLDERS=model ..
Make sure the files of the acoustic model are directly inside the HMM_FOLDERS
:
$ cd /path/to/models
$ ls *
model1:
feat.params mdef means sendump transition_matrices variances
model2:
feat.params mdef means sendump transition_matrices variances
You can do the same thing with statistical language models and dictionary files, using the following CMake parameters:
- Acoustic models:
HMM_BASE
andHMM_FOLDERS
, - Statistical language models:
LM_BASE
andLM_FILES
. - Dictionary files:
DICT_BASE
andDICT_FILES
.
For Emscripten parameters, see the wiki page on Github, which explain what they are and how they can affect pocketsphinx.js.
Please note that:
- If you want to package files, you need to set both
..._BASE
and..._FOLDERS
or..._FILES
. - If you do not specify an acoustic model to package, make sure you load one from external JavaScript files as described in a later section.
- By default, the first provided packaged acoustic model will be loaded if none is specified before the recognizer is initialized. If you attempt to initialize the recognizer without the
"-hmm"
parameter given in the config and no packaged acoustic model, initialization will fail. The model can be selected by giving the"-hmm"
parameter. See upcoming sections for how to specify recognizer parameters. - Make sure you optimize the size of your acoustic models (for instance, package
sendump
ormixture_weights
, not both, see PocketSphinx docs). - Statistical language models and dictionary files are optional at init time. As explained later, grammars and dictionary words can be added at runtime.
- If you want to package statistical language models, you must provide a dictionary that contains all words used in the SLMs.
- The PocketSphinx parameter for dictionary files is
"-dict"
and for language models"-lm"
. See next sections for how to specify recognizer parameters.
Unless you are using a small acoustic model and no large dictionary nor statistical language model, you would probably want to have these files packaged into separate JavaScript files, that should be loaded before pocketsphinx.js
. To do that, give the -DHMM_EMBED=OFF
option when running cmake to skip the embedding of the acoustic model files. You can still set HMM_BASE
and HMM_FOLDERS
which would be used to determine the default acoustic model to load.
For the dictionary and statistical language model, just ignore them when running cmake.
To generate the JavaScript files that contain these files, use emscripten's tools/file_packager.py
utility. For instance, to package the hub4
acoustic model provided with pocketsphinx
on cmusphinx
's subversion repository:
# cd .../cmusphinx/models/hmm/en_us
# python .../emscripten/tools/file_packager.py .../pocketsphinx.js/build/pocketsphinx.js --embed hub4wsj_sc_8k/mdef --js-output=mdef.js
# python .../emscripten/tools/file_packager.py .../pocketsphinx.js/build/pocketsphinx.js --embed hub4wsj_sc_8k/variances --js-output=variances.js
... and so on
Then, make sure you load all these generated JavaScript files (mdef.js
, variances.js
, etc.) before you load pocketsphinx.js
.
You can interact with pocketsphinx.js
directly if you need to, but it is probably easier to build your application against the API of recognizer.js
described in a later section.
The file pocketsphinx.js
can be directly included into an HTML file but as it is fairly large (a few MB, depending on the optimization level used during compilation and packaged files), downloading and loading it will take time and affect the UI thread. So, as explained later, you should use it inside a Web worker, for instance using recognizer.js
.
This API is based on embind
, you should probably have a look at that section in emscripten's docs to understand how to interact with emscripten-generated JavaScript. Earlier versions of Pocketsphinx.js used a C-style API which is now deprecated, but it is still available in the OBSOLETE_API
branch.
As a first example, to create a new recognizer:
var recognizer = new Module.Recognizer();
/* ... */
recognizer.delete();
Calls to pocketsphinx.js
functions are synchronous, that's also why you probably need to load it in a Web Worker, as explained in later sections.
Most calls return a ResultType object, which can be one of the following:
- SUCCESS, if the action was performed successfully.
- BAD_STATE, if the current state does not allow the action.
- BAD_ARGUMENT, if the argument provided is invalid.
- RUNTIME_ERROR, if there is a runtime error in the recognizer.
In JavaScript these values can be referred as Module.ReturnType.SUCCESS
, Module.ReturnType.BAD_STATE
, etc. For instance:
var recognizer = new Module.Recognizer();
/* ... */
if (recognizer.reInit(config) != Module.ReturnType.SUCCESS)
alert("Error while recognizer is re-initialized");
According to embind
's documentation, all objects created with the new
operator must be deleted explicitly with a .delete()
call.
The entry point of pocketsphinx.js
is the recognizer object. You can create as many instances as you want, but you probably don't need to and want so save memory. When a new instance is created, an optional Config
object can be given which will be used to set parameters used to initialize Pocketsphinx. Refer to Pocketsphinx documentation to learn about the possible parameters. A Config
object is basically an array of key-value pairs:
var config = new Module.Config();
config.push_back(["-fwdflat", "no"]);
var recognizer = new Module.recognizer(config);
config.delete();
/* ... */
recognizer.delete();
This will initialize a recognizer with "-fwdflat"
set to "no"
.
If you have included several acoustic models when compiling pocketsphinx.js
, you can select which one should be used by setting the "-hmm"
parameter. Say you have two models, one for English, one for French, and you have compiled the library with -DHMM_FOLDERS="english;french"
, you can initialize the recognizer with the French model by setting the correct value in the Config
object:
var config = new Module.Config();
config.push_back(["-hmm", "french"]);
var recognizer = new Module.recognizer(config);
If you do not give the "-hmm"
parameter, or give it an invalid value, the first model in the list will be used (here, english
).
Similarly, you should use recognizer config parameters to load a statistical language model ("-lm"
) or dictionary ("-dict"
) you have previously packaged inside pocketshinx.js
. Note that if you want to use a SLM, you must also have a dictionary file that contains the words used in the SLM.
In addition, a recognizer object can be re-initialized with new parameters after the instance was created, with a call to reInit
, for instance:
var config_english = new Module.Config();
config_english.push_back(["-hmm", "english"]);
var config_french = new Module.Config();
config_french.push_back(["-hmm", "french"]);
var recognizer = new Module.recognizer(config_english);
/* ... */
if (recognizer.reInit(config_french) != Module.ReturnType.SUCCESS)
alert("Error while recognizer is re-initialized");
Dictionary and language model files can be packaged at compile time as explained previously. Meanwhile, dictionary words, grammars (Finite State Grammars, FSG) or key phrases (for keyword spotting) can be added at runtime.
All words used in grammars or for keyword spotting must be present in the pronunciation dictionary. Refer to the CMU Pronunciation Dictionary site if you are not familiar with it. Words can be added as a vector of pairs word-pronunciation:
var recognizer = new Module.Recognizer();
var words = new Module.VectorWords();
words.push_back(["HELLO", "HH AH L OW"]);
words.push_back(["WORLD", "W ER L D"]);
if (recognizer.addWords(words) != Module.ReturnType.SUCCESS)
// Probably bad format used for pronunciation
alert("Error while adding words");
words.delete()
Note that PocketSphinx allows you to input several pronunciation alternatives for a word, by adding suffixes to it ((2)
, (3)
, etc.). However, adding a word with a suffix before the word without suffix will fail when calling addWords
:
words.push_back(["HELLO", "HH AH L OW"], ["HELLO(2)", "HH EH L OW"]); // OK
/* ... */
words.push_back(["HELLO", "HH AH L OW"], ["HELLO", "HH EH L OW"]); // Invalid
/* ... */
words.push_back(["HELLO(2)", "HH AH L OW"], ["HELLO", "HH EH L OW"]); // Invalid
There is also a call to check whether a word is in the dictionary. It returns its pronunciation if it is, or an empty string if it is not:
var recognizer = new Module.Recognizer();
var words = new Module.VectorWords();
words.push_back(["HELLO", "HH AH L OW"]);
recognizer.addWords(words);
var pronunciation = recognizer.lookupWord("HELLO"); // "HH AH L OW"
var empty = recognizer.lookupWord("GOODBYE"); // ""
A FSG is a structure that includes an initial state, a last state as well as a set of transitions between these states. Again, make sure all words used in transitions are in the dictionary (either loaded through a packaged dictionary file or added at runtime using addWords
). Here is an example of inputting one grammar:
var transitions = new Module.VectorTransitions();
// log-probability is 0 (i.e. probability is 1.0):
transitions.push_back({from: 0, to: 1, logp: 0, word: "HELLO"});
transitions.push_back({from: 1, to: 2, logp: 0, word: "WORLD"});
// null-transition:
transitions.push_back({from: 1, to: 2, logp: 0, word: ""});
var ids = new Module.Integers();
if (recognizer.addGrammar(ids,
{start: 1,
end: 2,
numStates: 3,
transitions: transitions}) != Module.ReturnType.SUCCESS)
alert("Error while adding grammar"); // Meaning that the grammar has issues
transitions.delete();
var id = ids.get(0); // This is the id assigned to the grammar
ids.delete();
Notice the Integers
object that is used to return an id back to the app to refer to the grammar. This id is then used to switch the recognizer to using that specific grammar. You will note that new Module.Integers()
actually returns a vector object that is then passed as a reference to addGrammar
. If the call is successful, the first element of the array is the id assigned to the grammar.
id
s usually start with 1
, 0
being kept for the default search, which is a language model, grammar file or key phrases file added at init time.
PocketSphinx also includes a keyword spotting search. Give the decoder a keyword or key phrase to catch and you can get, at any time, the number of times it was spotted. The key phrase is just a string with the phrase to spot. All words from the phrase must have been previously added with addWord
.
words.push_back(["HELLO", "HH AH L OW"], ["WORLD", "W ER L D"]);
recognizer.addWords(words);
var ids = new Module.Integers();
if (recognizer.addKeyword(ids, "HELLO WORLD") != Module.ReturnType.SUCCESS)
alert("Error while adding key phrase"); // Meaning that the key phrase has issues
var id = ids.get(0); // This is the id assigned to the search
ids.delete();
Note that there is a threshold that can be set to define how sensitive the search is. Add ["-kws_threshold", "1e-35"]
for instance to the config object. Values like "1e-50" mean that the keyword is more likely to be spotted but more likely to get false positives, while "1e-0" is restrictive and may miss actual keyword utterances. Experiment to find the ideal threshold. It varies greatly depending on the keyword itself, audio quality, and background noise.
{command: 'initialize', data: [["-kws_threshold", "1e-35"]]}
Note that you can also add key phrases via a file, using the "-kws"
argument as shown in the live.html
example.
A recognizer object can have any number of grammars and keyword searches but only one can be active at a time. The active search is the one used when there is a call to start()
, described later in this document. To switch to a specific search, you must use the id that was given during the call to addGrammar
or addKeyword
.
// id is the first element of the ids vector after call to addGrammar or addKeyword:
if (recognizer.switchSearch(id) != Module.ReturnType.SUCCESS)
alert("Error while switching search"); // The id is probably wrong
If you added a language model, grammar file, or key phrases file, the recognizer can switch back to it using id=0
.
To recognize audio, one must first call start
to initialize recognition, then feed the recognizer with audio data with calls to process
and finally call stop
once done. During and after recognition, the recognized string can be retrieved with a call to getHyp
.
Before calling start, one must make sure that the current language model is the correct one, mainly, whatever happened last:
- If a grammar or keyword search has just been given to the recognizer, it is automatically used as current language model.
- If a call to
switchSearch
was successful, the specified search will be used in the next call tostart
. - If a SLM was packaged in
pocketsphinx.js
and was loaded by being added in the parameters to theConfig
object used when the recognizer was instantiated (or re-initialized), then this model is the current language model.
Calls to process must include audio buffers in the form of an AudioBuffer
object. AudioBuffer
objects can be re-used. They must contain audio samples, as 2-byte integers, recorded at 16kHz (unless your acoustic model uses different characteristics).
Here is an example:
var array = ... // array that contains an audio buffer
var buffer = new Module.AudioBuffer();
for (var i = 0 ; i < array.length ; i++)
buffer.push_back(array[i]); // Feed the array with audio data
var output = recognizer.start(); // Starts recognition on current language model
output = recognizer.process(buffer); // Processes the buffer
var hyp = recognizer.getHyp(); // Gets the current recognized string (hypothesis)
/* ... */
for (var i = 0 ; i < array.length ; i++)
buffer.set(i, array[i]); // Feed buffer with new data
output = recognizer.process(buffer);
hyp = recognizer.getHyp();
/* ... */
output = recognizer.stop();
// Gets the final recognized string:
var final_hyp = recognizer.getHyp();
buffer.delete();
Remember to check the return values of the different calls and compare them to Module.ReturnType....
.
For a keyword spotting search, use addKeyword
instead of addGrammar
as explained previously. getHyp
returns as many times the keyphrase as it appeared since recognition started.
The recognition process also produces the segmentation, called hypseg in Sphinx jargon. It can be retrieved the same way as the hypothesis, with a getHypseg
call. It uses a Segmentation
object which is a vector of SegItem
objects, that contain the following fields: word
for the current recognized word, start
for the start frame of the word (one frame is 10ms), and end
for the end frame. The Segmentation is passed as a reference:
var seg = new Module.Segmentation();
if (recognizer.getHypseg(seg) == Module.ReturnType.SUCCESS) {
for (var i = 0 ; i < seg.size(); i++) {
var segItem = seg.get(i);
console.log("Word " + segItem.word +
" starts at frame " + segItem.start +
" ends at frame " + segItem.end);
}
}
seg.delete();
In most cases you probably don't need to do that, but to free the memory used by the recognizer, you must call recognizer.delete()
. Since you can re-initialize a recognizer with new parameters with a call to reInit
, this should be only necessary if you're sure you don't need any recognizer object anymore.
We have so far only dealt with words based on ASCII characters. We can also use unicode strings, but they must be manually encoded and decoded:
- Words added to
VectorWords
,VectorTransitions
or queried withlookupWord
must be encoded if they use non-ASCII characters. - Words retrieved in
Segmentation
object or fromgetHyp
calls must be decoded if they use non-ASCII characters. - There are unicode encoding and decoding functions in
recognizer.js
, you can re-use them and see where they are used. - If you use
pocketsphinx.js
viarecognizer.js
, you do not need to worry about encoding and decoding.
Using recognizer.js
, pocketsphinx.js
is downloaded and executed inside a Web worker. The file is located in webapp/js/
, both recognizer.js
and pocketsphinx.js
must be in the same folder at runtime. It is intended to be loaded as a new Web worker object:
var recognizer = new Worker("js/recognizer.js");
You can then interact with it using messages.
Messages posted to the recognizer worker might include the following attributes:
command
, command to be executed,data
, data to be passed to the command,callbackId
, id to be passed to the outgoing message, might be used to trigger a callback.
The worker sends messages back to the UI thread, either to call back when actions have been performed, report errors or send periodic information such as the current recognition hypothesis.
Messages posted by the recognizer worker might include:
status
, which can be eitherdone
orerror
,command
, the command that sent the message,code
, an error code,id
, a callback id that was given in the received incoming message,data
, additional data that the callback function might make use of,hyp
, the current recognition hypothesis,final
, a boolean that indicates whether the hypothesis is final (sent after call tostop
).
The error codes returned in messages posted back from the worker can be:
- the error code returned by
pocketsphinx.js
as explained previously, - or one of the following strings:
- "js-data", if the provided data are invalid,
- "js-no-recognizer", if the recognizer is not initialized.
Once the worker is created, the recognizer must be initialized:
// This value will be given in the message received after the action completes:
var id = 0;
recognizer.postMessage({command: 'initialize', callbackId: id});
Once it is done, the recognizer will post a message back, for instance:
{status: "done", command: "initialize", id: clbId}
, if successful, whereclbId
is the callback id given in the original command message.{status: "error", command: "initialize", code: initStatus}
, if there is an error, whereinitStatus
is the value returned by the call topsInitialize
, see above for possible values.
Recognizer parameters to be passed to PocketSphinx
can be given in the call to initialize
. For instance:
recognizer.postMessage({command: 'initialize',
callbackId: id,
data: [["-hmm", "french"],
["-fwdflat", "no"],
["-dict", "french.dic"],
["-lm", "french.DMP"]]
});
This will set the pocketsphinx
command-line parameter -fwdflat
to no
and initialize the recognizer with the acoustic model french
, the language model french.DMP
and the dictionary french.dic
, assuming pocketsphinx.js
was compiled with such models.
Note that once it is initialized, the recognizer can be re-initialized with different parameters. That way, for instance, a web application can switch between different acoustic and language models at runtime.
Words to be recognized must be added to the recognizer before they can be used in grammars. See previous sections to know more about the format of dictionary items. Words can be added at any time after the recognizer is initialized, and several words can be added at once:
// An array of pairs [word, pronunciation]:
var words = [["ONE", "W AH N"], ["TWO", "T UW"], ["THREE", "TH R IY"]];
recognizer.postMessage({command: 'addWords', data: words, callbackId: id});
The message back could be:
{id: clbId}
, the provided callback id, if given, as explained before, if successful.{status: "error", command: "addWords", code: code}
, if error, where possible values of the error code was described above.
Note that words can have several pronunciation alternatives as explained in Section 3.3.a.
As described previously, any number of grammars or keyword searches can be added. The recognizer can then switch between them.
A grammar can be added at once using a JavaScript object that contains the number of states, the first and last states, and an array of transitions, for instance:
var grammar = {numStates: 3,
start: 0,
end: 2,
transitions: [{from: 0, to: 1, word: "HELLO"},
{from: 1, to: 2, logp: 0, word: "WORLD"},
{from: 1, to: 2}]
};
recognizer.postMessage({command: 'addGrammar', data: grammar, callbackId: id});
All words must have been added previously using the addWords
command.
Notice that logp
is optional, it defaults to 0. word
is also optional, it defaults to ""
which is a null-transition.
In the message back, the grammar id assigned to the grammar is given. It can be used to switch to that grammar. So the message, if successful, would be like {id: clbId, data: id, status: "done", command: "addGrammar"}
, where id
is the id of the newly created grammar. In case of errors, the message would be as described previously.
Similarly, keyword spotting search can be added by just providing the key phrase to spot:
var keyphrase = "HELLO WORLD";
recognizer.postMessage({command: 'addKeyword', data: keyphrase, callbackId: id});
Just as like with grammars, words should already be in the recognizer, and the id of the newly added search is given in the callback. As explained previously, you might want to ajust the sensitivity threshold when initializing the recognizer, for example with providing ["-kws_threshold", "1e-35"]
.
The message to start recognition should include the id of the grammar (or keyword search) to be used:
// id is the id of a previously added grammar:
recognizer.postMessage({command: 'start', data: id});
Audio samples should be sent to the recognizer using the process
command:
// array is an array of audio samples:
recognizer.postMessage({command: 'process', data: array});
Audio samples should be 2-byte integers, at 16 kHz.
While data are processed, hypothesis will be sent back in a message in the form {hyp: "RECOGNIZED STRING"}
. If it is a keyword spotting search, the hyp
field will be the key phrase, present as many times as it appeared since recognition started.
Recognition can be simply stopped using the stop
command:
recognizer.postMessage({command: 'stop'});
It will then send a last message with the hypothesis, marked as final (which means that it is more accurate as it comes after a second pass that was triggered by the stop
command). It would look like: {hyp: "FINAL RECOGNIZED STRING", final: true}
.
The recognizer worker can load any file to make them available to pocketsphinx.js
. It can be an acoustic model, dictionary, language model or list of key phrases. There are two ways to do this:
- package the files into JavaScript files and load them,
- use raw (binary or text) files and lazy-load them.
The second solution is probably wiser as binary files packaged into JavaScript will end up much larger.
As shown in section 2.b.ii, Emscripten can package files inside JavaScript files so they can be accessed via the virtual file system. Use the load
command with the list of JavaScript files to load them. As for other calls in the worker, this will not affect the UI thread, and will call back once all files are loaded. The files packaged inside those JavaScript files are available when the recognizer is initialized or re-initialized with specific values of -hmm
, -lm
or -dict
.
recognizer.postMessage({command: 'load',
callbackId: id,
data: ["mymodel/mdef.js",
"mymodel/transition_matrices.js",
...
"mymodel/variances.js"]
});
The path are relative to the location of recognizer.js
, and in this example, the model can be loaded by using a config
object with ["-hmm", "mymodel"]
.
There will be an error callback with NETWORK_ERROR
if any of the files can't be loaded.
Emscripten provides a way to lazy-load files, actual HTTP request happen when files are accessed. We provide a way to add those files via recognizer.js
. Use the lazyLoad
command with data
giving the folders to create on the virtual file system, the files to add to the file system and the mapping:
recognizer.postMessage({command: 'lazyLoad',
callbackId: id,
data: {folders: [["/", "zh_broadcastnews_ptm256_8000"]],
files: [["/zh_broadcastnews_ptm256_8000", "means", "../zh_broadcastnews_ptm256_8000/means"],
["/zh_broadcastnews_ptm256_8000", "variances", "../zh_broadcastnews_ptm256_8000/variances"],
["/zh_broadcastnews_ptm256_8000", "transition_matrices", "../zh_broadcastnews_ptm256_8000/transition_matrices"],
["/zh_broadcastnews_ptm256_8000", "sendump", "../zh_broadcastnews_ptm256_8000/sendump"],
["/zh_broadcastnews_ptm256_8000", "mdef", "../zh_broadcastnews_ptm256_8000/mdef"],
["/zh_broadcastnews_ptm256_8000", "feat.params", "../zh_broadcastnews_ptm256_8000/feat.params"],
["/zh_broadcastnews_ptm256_8000", "mixture_weights", "../zh_broadcastnews_ptm256_8000/mixture_weights"],
["/zh_broadcastnews_ptm256_8000", "noisedict", "../zh_broadcastnews_ptm256_8000/noisedict"]]
}
});
folders
is an array of pairs where the second element is the name of the folder to create and the first element is where this folder should be created in.files
is an array of triplets. The first element is the folder on the virtual file system where the virtual file will be created, the second element is the name of the virtual file and the third element is the actual raw file to add, with a path relative the the location ofrecognizer.js
.
The example given above adds the Chinese acoustic model provided by CMU Sphinx. If the URL of recognizer.js
is https://example.com/pocketsphinx/js/recognizer.js
, URLs of the models' binary files are https://example.com/pocketsphinx/zh_broadcastnews_ptm256_8000/means
, etc. Then the model can be loaded with parameters ["-hmm", "zh_broadcastnews_ptm256_8000"]
. You can see an example of that in the attached live web app, with kws.txt
and kws.dict
.
In order to facilitate the interaction with the recognizer worker, we have made a simple utility that helps associate callbacks to be executed when the worker posts a message responding to a command you sent. You can find callbackManager.js
in webapp/js
.
To use it, first create a new instance of CallbackManager:
var callbackManager = new CallbackManager();
When you post a message to the recognizer worker and want to associate a callback function to it, you first add your callback function to the manager, which gives you a callback id in return:
recognizer.postMessage({command: 'addWords',
data: words,
callbackId: callbackManager.add(
function() {alert("Words added");})
});
In the onmessage
function of your worker, use the callback manager to check and trigger callback functions:
recognizer.onmessage = function(e) {
if (e.data.hasOwnProperty('id')) {
// If the message has an id field, it
// means that we might have a callback associated
var clb = callbackManager.get(e.data['id']);
var data = {};
// As mentioned previously, additional data can be passed to the callback
// such as the id of a newly added grammar
if(e.data.hasOwnProperty('data')) data = e.data.data;
if(clb) clb(data);
}
// Check for other message types here
};
Check live.html
in webapp
for more examples.
When a new worker is instantiated, it immediately returns a worker object, but the actual download of the JavaScript files might take some time, especially in our case where pocketsphinx.js
is fairly large. One way of detecting whether the files are fully downloaded and loaded is to post a first message right after it is instantiated and wait for a message back from the worker.
var recognizer;
function spawnWorker(workerurl, onReady) {
recognizer = new Worker(workerurl);
recognizer.onmessage = function(event) {
// onReady will be called when there is a message
// back
onReady(recognizer);
};
recognizer.postMessage('');
};
The first message posted to the recognizer can include the name of the PocketSphinx JavaScript file to load. This is handy if you want to build an application with several different models, you can keep the same recognizer.js
file for different parts of your application and load any PocketSphinx JavaScript file that you want. By default, it will load pocketsphinx.js
, but if you want your application to load a file called pocketsphinx_chinese.js
, you can just add it as parameter to the first posted message:
var recognizer;
function spawnWorker(workerurl, onReady) {
recognizer = new Worker(workerurl);
recognizer.onmessage = function(event) {
// onReady will be called when there is a message
// back
onReady(recognizer);
};
recognizer.postMessage('pocketsphinx_chinese.js');
};
After the first message back was received, proper listening to onmessage can be added:
spawnWorker("js/recognizer.js", function(worker) {
worker.onmessage = function(e) {
// Add what you want to do with messages back from the worker
};
// Here is a good place to send the 'initialize' command to the recognizer
});
Of course, the worker must be able to respond to the first message, as we did in recognizer.js
:
function startup(onMessage) {
self.onmessage = function(event) {
self.onmessage = onMessage;
self.postMessage({});
}
};
// This function is called first, it triggers
// a first postmessage, then adds the proper respond to
// commands:
startup(function(event) {
switch(event.data.command){
//We deal with commands properly
}
});
All these are illustrated in webapp/live.html
and recognizer.js
.
We include an audio recording library based on the Web Audio API that accesses the microphone, gets audio samples, converts them to the proper sample rate (16kHz for our default acoustic model), and sends them to the recognizer. This library is derived from Recorderjs. To know more about audio capture and playback on the web, you could have a look at this overview of audio on the Web. A more complete documentation of the recorder can be found in doc/AudioRecorder/README.md
.
Include audioRecorder.js
in the HTML file and make sure audioRecorderWorker.js
is in the same folder. To use it, create a new instance of AudioRecorder
giving it as argument a MediaStreamSource
. As of Today, the Google Chrome and Firefox (25+) implement it. You also need to set the recognizer attribute to a Recognizer worker, as described above.
// Deal with prefixed APIs
window.AudioContext = window.AudioContext || window.webkitAudioContext;
navigator.getUserMedia = navigator.getUserMedia ||
navigator.webkitGetUserMedia ||
navigator.mozGetUserMedia;
// Instantiating AudioContext
try {
var audioContext = new AudioContext();
} catch (e) {
console.log("Error initializing Web Audio");
}
var recorder;
// Callback once the user authorizes access to the microphone:
function startUserMedia(stream) {
var input = audioContext.createMediaStreamSource(stream);
recorder = new AudioRecorder(input);
// We can, for instance, add a recognizer as consumer
if (recognizer) recorder.consumers.push(recognizer);
};
// Actually call getUserMedia
if (navigator.getUserMedia)
navigator.getUserMedia({audio: true},
startUserMedia,
function(e) {
console.log("No live audio input in this browser");
});
else console.log("No web audio support in this browser");
Once the recorder is up and running, you can start and stop recording and recognition with:
// To start recording:
recorder.start();
// The hypothesis is periodically sent by the recognizer, as described previously
// To stop recording:
recorder.stop(); // The final hypothesis is sent
The constructor for AudioRecorder can take an optional config object. This config can include a callback function which is executed when there is an error during recording. As of today, the only possible error is when the input samples are silent. It can also include the output sample rate, which you might need to set if you use an acoustic model of 8kHz audio.
var audioRecorderConfig = {
errorCallback: function(x) {alert("Error from recorder: " + x);},
outputSampleRate: 8000
};
recorder = new AudioRecorder(input, audioRecorderConfig);
All these are illustrated in the given live demo, in the webapp/
folder.
Note that live audio capture is only available on recent versions of Google Chrome and Firefox. Chrome, prior to version 29, only produced silent audio on many platforms. Firefox includes the necessary features starting from version 25.
The file webapp/live.html
is an example of live recognition using the web audio API. It works on Chrome and Firefox (25+), if the web audio API actually works. Note that we observed the recorded audio to be silent on some configurations we have tried.
To build an application, this is a good starting point as it illustrates the different components described in this document. In that demo, three different grammars are available and the app can switch between them.
There is also:
live_kws.html
for keyword spotting,live_zh.html
for recognition of Chinese.
There is a test suite in tests/js
which makes use of QUnit. There is a README file inside the folder.
If you are not familiar with speech recognition, you might need to take some time to learn some of the concepts, mainly:
- acoustic models (we provide one small model for English but other Sphinx acoustic models can be used as well),
- language models (grammars (FSG, JSGF), statistical language models (n-grams)),
- Cepstral Mean Normalization (CMN) and the different CMN strategies.
In terms of performance, you should get exactly the same result as using PocketSphinx compiled on other platforms. For instance, because of the CMN policy, the accuracy of the first utterance is usually pretty bad, especially for non-native speakers.
The am
folder contains an acoustic model trained with SphinxTrain. It is built using the RM1 corpus, semi-continuous, with 200 senones.
PocketSphinx.js now uses PocketSphinx (and Sphinxbase) code as it is in its github repository, using git submodules.
PocketSphinx licensing terms are included in the pocketsphinx
and sphinxbase
folders.
The files webapp/js/audioRecorder.js
and webapp/js/audioRecorderWorker.js
are based on Recorder.js, which is under the MIT license (Copyright © 2013 Matt Diamond).
The remaining of this software is licensed under the MIT license:
Copyright © 2013-2017 Sylvain Chevalier
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.